Related Experiment Videos
Econometric mixture models and more general models for unobservables in duration analysis
1Department of Economics, University of Chicago, IL 60637.
Statistical Methods in Medical Research
|January 1, 1994
Summary
This study examines unobserved factors in duration models, showing how data variations aid model identification. It highlights limitations in traditional methods, even when models appear identifiable but are hard to estimate.
Area of Science:
- Econometrics
- Survival Analysis
- Statistical Modeling
Background:
- Duration models are crucial for analyzing time-to-event data.
- Unobserved heterogeneity presents challenges in duration model estimation.
- Existing identification strategies may not guarantee consistent estimation.
Purpose of the Study:
- To investigate the role of unobservables in duration models.
- To demonstrate how cross-section and time-series variation improve model identification.
- To assess the limitations of traditional identification approaches.
Main Methods:
- Utilizing cross-section and time-series variation in regressors.
- Applying identification strategies to single-spell, competing risks, and multiple spell duration models.
- Analyzing a specific case where conventional identification fails consistent estimation.
Main Results:
- Cross-section and time-series variation effectively facilitate the identification of various duration models.
- Traditional identification methods can be insufficient for consistent estimation.
- The study provides a counterexample to conventional identification assumptions.
Conclusions:
- Data variation is key for robust duration model identification.
- Careful consideration of estimation challenges beyond conventional identification is necessary.
- The findings offer practical insights for applied econometricians and statisticians.